Fantastic Futures: Artificial intelligence and Cultural Heritage

What is the impact of artificial intelligence on GLAM (or digital cultural heritage) collecting and practice?  

It is worth retracing the theme set for the plenary sessions from the 2019 Fantastic Futures conference on AI for Libraries, Archives, and Museums held at Stanford University recently to frame more detailed questions and get closer to finding some answers: 

  • How will AI transform the library?  How will the library transform AI?
  • What can we do with AI?  What should we do?  
See through the zeros and ones.
See through the zeros and ones.

If you’re asking these questions (like I am) about artificial intelligence and how machine learning works, and you are a knowledge worker with a professional background in research or cultural heritage, then it might be worth taking time to listen to the video of the plenary talks from the first two Fantastic Futures conferences (2018 and 2019).  

For some, there’s a lot of reading to do, especially if you’re interested in: 

For others tackling the whole idea of AI as a subject, then it might be heading into a free online course on the basics of artificial intelligence created by the:  Fantastic Finns!

The takeaways from this conference for me were:

  • Technical: workshops and technique and tool sharing approaches are vital!  On minimal coding ability with Python and familiarity with how Jupyter notebooks work, doing the Intro to TensorFlow workshop in Colab (Colaboratory is a Google service that supports Python 2.7 and 3.6) opened up a world of possibility. But, that is because all of the steps were laid out and there was an explanatory component . So some further work ahead for those of us keen to become more technically proficient to unpack the steps and the commands to aid with building coding practice.  
  • Community: collaboration and coordination is needed, especially around any development of co-created reference datasets, and the use of and or recommendation of appropriate datasets and models.  
  • Curation: strategy and ethics (rather than technology) needs to drive the agenda to integrate machine learning into heritage practices, e.g. setting terms and conditions for use of collections, access arrangements to vendor controlled content, and social responsibility. Read a great piece by Ben Green ‘Good’ isn’t good enough – that talks about putting the Algorithm In The Loop (a reversal of the Human In the Loop when it comes to algorithm development).     

Next steps? 

In Australia, look into the AI Collaborative Network.  

Join the Artificial Intelligence for Libraries, Archives, and Museums (aka AI4LAM) conversation online: